Status of Initiating Pattern of Hemodialysis: A Multi-center Study

Placement of a vascular access is essential in patients who receive maintenance hemodialysis (HD). National Kidney Foundation-Dialysis Outcomes Quality Initiative Clinical Practice Guidelines for Vascular Access recommend placement of a permanent vascular access (arteriovenous fistulae [AVF] or arteriovenous grafts [AVG]) before initiation of maintenance HD to avoid the use of central venous catheters (CVC) (1). This is because permanent accesses provide better blood flow rates (2) and are associated with less complication than CVC (3-6). In addition, unplanned initiation of HD is associated with poorer outcomes than planned initiation (7). However, according to previous reports, roughly two thirds of patients initiate HD with CVC (8-10), although the use of permanent access increases over time (8), and a majority of patients go on unplanned HD at initiation (11-13).

For prior placement and use of a permanent access at initiation of HD, time is required for patient education, selection of dialysis modality, selection of the access type, surgical creation and maturation of the access. In this context, adequate time for predialysis care is essential. Previous studies have found that late referral to a nephrologist is a major determinant for use of CVC at initiation (8, 9, 14, 15) and is associated with poorer outcomes than early referral (7). We also found that late referral increased the risk of cardiovascular disease-related mortality in HD patients (16). Therefore early referral is crucial for reducing the use of CVC and unplanned initiation of HD and improving outcomes.

Although there are a few reports of predialysis care and initial pattern of HD in Korea (14, 15), they were data from a single center. In addition, Asian data about the type of access in use and mode of presentation (unplanned vs. planned) in patients initiating HD are scarce. According to the database of the end-stage renal disease registry committee in Korean Society of Nephrology, the prevalence of end-stage renal disease (ESRD) is increasing rapidly, and the number of HD patients accounts for three fourths of patients on dialysis therapy (17). Thus we undertook a multi-center study to investigate the type of access, presentation mode of the first HD and the significance of care by a nephrologist in initial pattern of HD and 6 month-mortality.

Our study shows the discrepancy between clinical practice and recommendations by guidelines in initiation of maintenance HD. Most patients initiated HD in an unplanned manner, without prior creation of a permanent access, and used CVC as their first access. More patients who received at least 3 months of care by a nephrologist started HD in a planned manner, with prior creation of a permanent access, and more used a permanent access than those referred later. Nevertheless, only a minority of patients with predialysis care initiated planned HD with permanent accesses. This is the first multi-center report of how patients prepare and start HD in Asia including Korea, and it is not just a problem of a facility as it was conducted from a cohort of 8 hospitals in 4 provinces. Furthermore, the results demonstrate the clinical implication of predialysis care in preparing HD in practice.

Although the use of permanent access increases over time (8), most patients use CVC at initiation of HD. Lenz et al. (18) reported the incidence of initial CVC use to be 92.3%, Astor et al. (10) and Stehman-Breen et al. (9) to be about 66%. In our cohort, the proportion of patients who started HD in an unplanned manner (81.9%) and that of patients who used CVC as their first access (86.3%) were both higher than other reports (7, 9, 10). The higher frequency may be explained by two reasons. One reason is the difference in population. While most studies excluded patients who were to enter PD therapy and who were to receive or who previously received renal transplantation, there were 82 patients (16.3%) in those categories included in our study. We included patients who changed to PD or renal transplantation because many patients start HD without choosing their dialysis modality in clinical practice. The second reason is that the definition of planned initiation of HD differed from other reports. Lorenzo et al. (7) defined planned initiation as more than 3 months of care by a nephrologist before starting HD, and Marrón et al. (12) defined it as outpatient, scheduled initiation. Our definition of planned initiation was starting HD with prior placement of a permanent access, regardless of whether the access is functional. This was because we considered the creation of a permanent access is the most crucial and difficult step in planning and preparing HD. Other multiple reasons may be attributable, such as center-specific guidelines, coexisting diseases, and hemodynamic status. However, even considering the difference, this higher rate reflects that our clinical practice is far from what guidelines recommend, which was consistent with other studies (7-13, 18, 21).

As shown in Table 4, the most common reasons for CVC use were patient-related factors. This shows that the patient's reluctance is the major barrier of access planning in our clinical practice. Our result is different from results by Lenz et al. (16), in which the most common reasons were inadequate predialysis care. There are three explanations for the higher proportion of patient-related factors than the absence of predialysis care in our results. First, the definition of patientrelated factors was less strict than that of other study (18). In our study, no show was defined as failure to attend scheduled clinic appointments for at least 3 months, in contrast other study (18) defined it as failure to attend at least half of scheduled appointments. Second, 'absence of predialysis care' was judged to be responsible only when the patient had no obvious reasons other than the absence of predialysis care. Third, the proportion of patients who refused surgery for vascular access was higher in our cohort than in the cohort by Lenz et al. (18). In our result, refusal of surgery accounted for 34.8% of total CVC users (n=434). In contrast, in the cohort by Lenz et al. (18), it accounted for only 1.9% of total CVC users (n=157).

The presence of predialysis care was an important factor in determining both the mode of presentation and initial access type. Thus early referral to a nephrologist is indeed essential for increasing numbers of patients initiating planned HD and using permanent accesses (8, 9, 11-15, 18, 22). However, there are two considerable aspects in our study. One is the quality of predialysis care. It is remarkable that despite more than 3 months of care by a nephrologist, most patients started unplanned HD (75.9%) and used CVC initially (81.1%). In our study, the proportion of predialysis care was similar between centers. But the rate of planned HD and use of permanent access differed between centers, ranging from 4% to 43% (Table 3). This indicates that the quality of predialysis care differed between centers. Dialysis education by nephrologists and HD nurses and frequent medical visits can influence the timing and planning of HD (11, 12, 23), and efforts of physicians for vein preservation and preoperative vein mapping can influence the choice of access type (24). Thus a multidisciplinary team-based care is suggested to improve the quality of predialysis care, which incorporates nephrologists, nurses, dieticians and social workers (19, 25, 26). Another aspect is the definition of early referral. As the economical burden of the insurance system and concept of dialysis differ between countries, the accurate time for early referral is difficult to estimate. However, we should consider factors that are out of our control, such as acute need for dialysis, rapid GFR loss or poor maturation of the access, which accounted a substantial portion of reasons for CVC use in our study (Table 4). Buck et al. (11) reported that service-related factors were responsible for urgent initiation of HD among patients who had been seen by a nephrologist at least 4 months, including delayed discussion and counseling by HD nurses and late referral for access creation. Lenz et al. (18) found that there were patient-related factors, physician-related factors and unexpected reasons for CVC use despite at least 4 months of predialysis care. To overcome these problems, 3-4 months may be insufficient for adequate predialysis care, and from our results, the optimal period for predialysis care seems to be more than 3 months. Some authors suggested it has to be at least 12 months (11, 27, 28). Although a statistical significance was not achieved, patients with predialysis care at least 12 months tended to use more permanent accesses at initiation than patients with predialysis care less than 12 months in this study (Fig. 3). To determine the adequate length of time for predialysis care, further study is required.

The time of permanent access placement is also of importance in determining the access type at initiation. In our study, about a quarter of patients who had prior access placement used CVC as their first access, and the interval between access placement and initial HD was significantly shorter in CVC users than in permanent access users (3.4±3.9 weeks vs. 16.8±20.8 weeks). This indicates that delay in creating a permanent access was a significant contributor for CVC use even in patients who had prior access placement. As shown in Table 4, acute need for dialysis (22.7%), poor maturation of the access (22.7%), delay in referral to a surgeon (22.7%) and delay between referral and surgery (22.7%) accounted most reasons for CVC use. In addition, of patients who did not have prior access placement, one third used CVC because of acute need for dialysis (26.0%), delay in referral to a sursurgeon (2.9%) and delay between referral and surgery (1.2%). If delays in access creation had been reduced, the frequent CVC use could have been reduced in 32.0% of the entire CVC users.

The encouraging part in our study is that, although a minority of patients used permanent accesses at their first HD, most (84.6%) changed to permanent accesses after 6 months and that AVF was the leading type in use (72%). This rate is higher than that of United States or Canada (8, 19, 21, 29), but lower than those of several countries in Europe (21). This indicates that a facility's preference and approaches to choice of access differ between countries and determine the access type in use. Therefore efforts to meet the goal for CVC use to <10% (1) can further increase the AVF use.

Contrary to our expectation, neither the absence of predialysis care, unplanned initiation, nor initial CVC use was a risk predicting deaths within 6 months (Table 5). This result differs from other reports which show the independent contribution of care by a nephrologist (8, 22) and the type of access (7, 10, 27, 30-32) in determining the patient outcome. Our result can be explained by two reasons. First, the follow-up period of 6 months may have been relatively short to assess the long-term survival. If the follow-up period were extended longer, there may be difference in mortality between the access types as other studies (7, 8, 15). Second, the overuse and imprudent preference to CVC of nephrologists ourselves may have made no difference in mortality rate.

In conclusion, most patients initiated unplanned HD and used CVC initially, in spite of predialysis care. To reduce the use of CVC in patients initiating HD, early referral for predialysis care, access planning, and timely attempts for creation of a permanent access is essential.